Rehabilitative Orbital Decompression for Graves' Orbitopathy: Risk Factors Influencing the New Onset of Diplopia in Primary Gaze,Outcome,and Patients' Satisfaction

Abstract

Background:

Patients with moderate to severe Graves' orbitopathy (GO) rather frequently require rehabilitative surgery after medical therapy. Diplopia is the most common side effect of orbital decompression (OD). The aim of this study was to evaluate the occurrence of postoperative diplopia in primary gaze after OD, and the influence of the surgical approach on this outcome. Moreover, we investigated the results in terms of proptosis reduction, and the long-term subjective satisfaction of patients treated with OD with regard to their appearance and ocular function.

Methods:

A retrospective evaluation of 247 patients with GO treated with medial and lateral decompression (MLD) or lateral decompression (LD) OD between January 2002 and December 2009 was performed.

Results:

The overall prevalence of postoperative diplopia in primary gaze was 55/247 (22.3%), with a statistically significant difference (p<0.001) between patients with (36/113, 31.2%) and those without (19/134, 14.2%) preoperative diplopia in secondary gaze. The surgical procedure influenced the outcome in patients without preoperative diplopia (17.8% after MLD and 0% after LD, p=0.02), but not in patients with preoperative diplopia in secondary gaze (33.3% after MLD and 26.1% after LD, p=0.5). Overall, proptosis reduction was 5.7±2.2 mm (1–11 mm), after MLD and 4.0±1.6 mm (1–8 mm) after LD (p<0.001). Fifty-one out of 55 patients with constant, postoperative diplopia in primary gaze after OD underwent squint surgery, which was successful in all but two. Four patients refused squint surgery. Patients were also interviewed for satisfaction in terms of recovery of their appearance and ocular function after a mean of 6 years from surgery (range 2–9 years): more than 85% of patients reported a good to excellent postoperative satisfaction for both items.

Conclusions:

Preoperative diplopia in secondary gaze is a risk factor for the development of diplopia in primary gaze after OD, independently of the surgical approach (MLD vs. LD). In absence of diplopia, MLD, but not LD, seems to be associated with its development in primary gaze. The reduction in proptosis after MLD is greater than that after LD. Most patients were satisfied with the results of both appearance and ocular function after OD.

Introduction

G raves' orbitopathy (GO) is the most common autoimmune inflammatory disorder of the orbit in adults (1,2). GO is characterized by an increase in the orbital content due to fibroblast proliferation, glycosaminoglycan production, edema, and muscle enlargement, ultimately causing an increased orbital pressure determining exophthalmos (proptosis) of variable degree. Due to its inflammatory features and muscle involvement, in addition to proptosis, the main clinical manifestations of GO are eyelid edema, conjunctival hyperemia, photophobia, impaired muscle function, and consequent diplopia and/or strabismus. In rare cases, sight impairment can occur, due to compression of the optic nerves at the orbital apex (3,4).

Treatment of GO includes systemic antinflammatory and immunosuppressive measures based on the use of glucocorticoids and orbital radiotherapy, as well as several surgical procedures, namely orbital decompression (OD), muscle surgery, and eyelid surgery (3,4). Urgent indications for OD are optic neuropathy, severe proptosis with consequent corneal exposure, and/or risk for subluxation of the eye globes. More commonly, OD can be performed for rehabilitative purposes (3). The rather common use of orbital surgery in GO for cosmetic reasons implies the need of refining the surgical techniques to reduce its complications (3).

Various techniques of OD are currently in use. Removal of the medial wall and of the orbital floor through a transantral approach is quite common (5,6). Starting from the early 1990s, a transnasal endoscopic technique was developed and became popular (7). In addition, removal of the lateral wall, either alone or in association with the medial wall, has been proposed by many authors (8 –12). Finally, Olivari has proposed the simple removal of intraorbital fat without osteotomies (13).

In our institution, the surgical approach is personalized for each patient depending on the degree of proptosis, the presence of optic nerve compression, and the presence of a significant restriction of one or more extraocular muscles. As a general principle, medial decompression (MD) is reserved for patients with mild proptosis (<23 mm) and without restriction of the medial rectus muscle, lateral decompression (LD) is used in patients with a moderate proptosis (23–26 mm), with or without extraocular muscle impairment, and balanced medial and lateral decompression (MLD) is used in case of more severe proptosis (>26 mm). In patients with optic neuropathy, the medial wall is always removed, with or without lateral wall removal, depending on the severity of the proptosis. Removal of the orbital floor, in addition to the balanced decompression, is performed only in particular situations such as very severe proptosis (>30 mm) with or without optic nerve compression and/or risk of ocular subluxation. Thus, the most widely used technique is MLD, followed by LD, whereas MD is used less frequently.

Although uncommon, complications of OD can be severe and permanent. They include dural damage with cerebrospinal fluid leak and/or pneumocephalus, iatrogenic damage to the optic nerve, intraorbital haemorrhage (14), temporary or permanent damage of the infraorbital and (more rarely) of supraorbital nerves, and worsening of pre-existent or new onset of diplopia (11). The last is the most common side effect, and its relation with the type of surgical technique has not yet been established. Postoperative diplopia in primary gaze can significantly lessen the quality of life of patients. Risk factors for postoperative diplopia (including diplopia in primary gaze) have not been clarified.

The aim of this retrospective study was to evaluate the occurrence of postoperative diplopia in primary gaze after OD, and to identify putative risk factors responsible for this complication. In addition, we investigated the results in terms of proptosis reduction, and the long-term subjective satisfaction of patients treated with OD with regard to their appearance and ocular function. For this purpose, we retrospectively studied a large series of patients with GO who had undergone MLD or LD between 2002 and 2009.

Materials and Methods

Patients

The medical records of 451 consecutive GO patients who had undergone OD between 2002 and 2009 were reviewed. All patients were operated by a single surgeon (S.S.F.). Inclusion criteria were: patients with inactive disease operated on by MLD or LD with rehabilitative purposes. Therefore, surgical indications included lagophthalmos and/or corneal exposure, proptosis with facial disfigurement, and the patient's choice for cosmetic and psychosocial reasons. Exclusion criteria were urgent surgery in active disease (optic neuropathy, subluxation of the eye), visual loss (because diplopia was not assessable), preoperative diplopia in primary gaze, postoperative follow-up shorter than 3 months, and incomplete preoperative or postoperative data. Thus, 56 patients were excluded, because they underwent OD using other techniques or as a revision procedure, 86 because of preoperative diplopia in primary gaze, 22 because of insufficient data (visual loss, inadequate follow-up), and 40 because surgery was performed to treat complications of the active disease (36 optic neuropathy and 4 subluxation of the eye). The remaining 247 patients (197 treated with MLD and 50 treated with LD) were eligible for the study. All patients had been previously treated with systemic (oral or IV) glucocorticoid therapy, alone or combined with orbital radiotherapy, and they were euthyroid at the moment of OD.

Patients underwent a preoperative and a postoperative ophthalmological evaluation, including Hertel exophthalmometry and/or computed tomography scan of the orbits, an analysis of extraocular muscle movements with arc perimeter, and a subjective semi-quantitative evaluation of diplopia, which was ranked as follows: absent; present in peripheral gaze (intermittent and inconstant); and present in primary gaze (constant). The following data were collected from the medical records: demographical data, surgical technique, preoperative and postoperative values of proptosis, preoperative extraocular muscle impairment (evaluation of diplopia in secondary gaze and ocular versions), visual function, worsening or new onset of postoperative diplopia in primary gaze, and surgical complications.

Patients were also interviewed by a telephone call to verify their satisfaction in terms of recovery of facial appearance and ocular function, after OD. One hundred and ninety-two/247 patients were reached by a telephone call (mean follow-up was 6 years, range 2–9 years).

The study was approved by our Institutional Review Board. Subjects gave their consent to interview and clinical data review.

Surgical techniques

All surgeries were performed under general anesthesia. Perioperative intravenous antibiotics were administered for at least 3 days in all patients.

Lateral wall OD

Through an upper blepharoplasty incision, a subperiosteal dissection of the lateral wall and of the outer portion of the superior and inferior walls was carried out. With the help of a malleable retractor, the eye was gently displaced, and with cutting and diamond burrs, the lateral wall of the orbit was drilled away, sparing the lateral orbital rim. The outer part of the orbital roof and zygomatic bone lateral to the infraorbital bundle were also removed. Then, the periorbita was incised from back to front to allow the dislocation of the orbital fat and, when feasible, a variable amount of fat was removed. A detailed description of this approach was previously reported (12).

Medial and lateral wall OD

The medial stage of the decompression was performed first to reduce pressure on the eye during the subsequent lateral approach. After decongestion of the nose with 1:200,000 epinephrine solution, transnasal ethmoidectomy, sphenoidotomy, and middle antrostomy were performed. Once the papiracea was skeletonized, it was carefully fractured with a Freer elevator or a curette, and its posterior two-thirds were removed piecemeal. The anterior third of the papiracea was spared in order to reduce the risk of postoperative diplopia as well as the obliteration of the naso-frontal recess. The orbital floor medial to the infraorbital nerve was removed only in its posterior third, with the help of a diamond burr if the bone was thick. The anterior two-thirds were spared to reduce the risk of infero-medial shifting of the orbital contents. As described, the lateral wall decompression was then performed. Finally, the periorbita was incised, both laterally and medially, to allow the dislocation of the intraorbital fat (11).

Endpoints

The primary endpoints of the study were the comparison of incidence of postoperative diplopia in primary gaze between MLD and LD and the identification of putative risk factors for the occurrence of the same postoperative complication. The secondary endpoints were the comparison of the two surgical techniques in terms of proptosis reduction and patient satisfaction in terms of recovery of their facial appearance, ocular function, and other side effects.

Statistical analysis

Statistics were performed using the STATA 9.1 software (College Station, TX). Comparisons between continuous data obtained before and after OD were performed by a paired t-test. The occurrence of diplopia after OD among different group of patients was analyzed with the chi-squared test. Risk factors for postoperative diplopia in primary gaze were analyzed by uni- and multivariate logistic regression models using a forward selection algorithm. Data are presented as mean±standard deviation (SD).

Results

Demographic data and surgical indications

Of the 247 patients, were 63 men and 184 were women. Mean age at the time of surgery was similar in bothsexes: 42.1±12.4 years (range 15–69) in men and 44.8±11.7 (range 17–79) in women.

Overall prevalence of postoperative diplopia

Preoperative diplopia in secondary gaze was present in 113 patients (45.7%) and was not statistically different between men (53.9%) and women (42.9%) (p=0.13). The overall prevalence of postoperative diplopia in primary gaze was 55/247 (total 22.3%, 27% men and 20.6% women, p=0.29). The prevalence of postoperative diplopia in primary gaze was significantly lower in patients without preoperative diplopia (19/134, 14.2%) than in those with preoperative diplopia (36/113, 31.2%) in secondary gaze (p<0.001) (Table 1).

Table 1.

Postoperative Primary Gaze Diplopia According to the Preoperative Secondary Gaze Diplopia

Preoperative status	n	No postoperative primary gaze diplopia	Postoperative primary gaze diplopia
No diplopia	134	115/134	19/134 (14.2%)^a
Secondary gaze diplopia	113	77/113	36/113 (31.2%)^a
Total	247	192/247	55/247 (22.3%)

p<0.001.

We found that 49/197 patients (24.9%) treated with MLD, and 6/50 patients (12.0%) treated with LD, experienced postoperative diplopia in primary gaze (p=0.05). The prevalence of postoperative diplopia in primary gaze after MLD was 19/107 (17.8%) in patients without preoperative diplopia and 30/90 (33.3%) in those with preoperative secondary gaze diplopia (p=0.01). In patients decompressed only laterally, the prevalence of postoperative primary gaze diplopia was 0/27 (0%) in those without preoperative diplopia, and 6/23 (26.1%) in those with preoperative diplopia in secondary gaze (p=0.005). In patients without preoperative diplopia in secondary gaze, we observed a statistically significant prevalence of a new onset of diplopia in the primary gaze after MLD (p=0.02). Among patients with preoperative diplopia in secondary gaze, we did not observe any statistically significant difference in terms of diplopia in the primary gaze when comparing the two surgical techniques (p=0.5) (Table 2).

Table 2.

Postoperative Primary Gaze Diplopia According to the Presence of Preoperative Secondary Gaze Diplopia and According to the Surgical Technique Used

	Postoperative primary gaze diplopia
Preoperative diplopia	Absent	Present	Total
MLD
No	88	19 (17.8%)^a,b	107
Secondary gaze	60	30 (33.3%)^a,c	90
Total	148	49 (24.9%)^d	197
LD
No	27	0 (0%)^b,e	27
Secondary gaze	17	6 (26.1%)^c,e	23
Total	44	6 (12.0%)^d	50

p = 0.01, MLD (No) vs. MLD (SG).

p = 0.02, MLD (No) vs. LD (No).

p = 0.5, MLD (SG) vs. LD (SG).

p = 0.05, overall prevalence of postoperative primary gaze diplopia in MLD vs. LD treatment.

p = 0.005, LD (No) vs. LD (SG).

LD, lateral decompression; MLD, medial and lateral decompression; (No), Patients with no preoperative diplopia; (SG), patients with preoperative secondary gaze diplopia.

Fifty-one out of 55 patients with constant, postoperative diplopia in primary gaze after OD underwent squint surgery, which was successful in all but two. Four out of 55 patients refused squint surgery (two patients for each group).

Risk factors affecting the onset of diplopia in primary gaze after OD

Univariate analysis showed that preoperative diplopia in secondary gaze, age at surgery, and impaired elevation of the eye (ocular versions in the upgaze) influenced the risk of postoperative diplopia in primary gaze (Table 3). Multivariate analysis showed that preoperative diplopia was the only parameter that influenced the incidence of postoperative diplopia in the primary gaze; whereas sex, age, year of surgery, surgical technique, severity of preoperative proptosis, ocular versions, extent of decompression (evaluated as proptosis reduction), and asymmetric presentation of the disease (evaluated as the difference in proptosis between the two eyes) did not correlate with the risk of postoperative diplopia in primary gaze.

Table 3.

Univariate Analysis of Putative Factors Responsible for a Higher Risk of Postoperative Primary Gaze Diplopia

Factors affecting onset	Odds ratio	p>\|z\|	[95% confidence interval]
Preoperative diplopia	2.82	0.001	[1.51–5.30]
Sex	0.70	0.30	[0.36–1.36]
Age at surgery	1.03	0.02	[1.00–1.05]
Preoperative proptosis	1.00	0.93	[0.88–1.13]
Difference in proptosis between the two eyes	1.21	0.19	[0.94–1.56]
Proposis reduction	1.01	0.87	[0.87–1.16]
Surgical technique	0.41	0.06	[0.16–1.02]
Year of surgery (balanced decompression)	0.92	0.16	[0.80–1.03]
Year of surgery (lateral decompression)	0.73	0.30	[0.40–1.35]
MRM version	1.04	0.38	[0.96–1.13]
LRM version	0.98	0.54	[0.92–1.04]
IRM version	0.94	0.45	[0.80–1.10]
SOM version	0.95	0.52	[0.81–1.11]
IOM version	0.93	<0.001	[0.89–0.96]
SRM version	0.93	<0.001	[0.89–0.96]

Statistically significant p-values (p<0.05) are presented in boldface.

IOM, inferior oblique muscle; IRM, inferior rectus muscle; LRM, lateral rectus muscle; MRM, medial rectus muscle; SOM, superior oblique muscle; SRM, superior rectus muscle.

Changes in proptosis after OD

Overall, preoperative and postoperative proptosis values were 25.0±2.4 mm (18–32 mm) and 19.7±2.5 mm (14–26 mm), respectively. Overall, proptosis reduction was 5.7±2.2 mm (1–11 mm) after MLD and 4.0±1.6 mm (1–8 mm) after LD (p<0.001).

Preoperative proptosis was statistically greater in patients undergoing MLD compared with patients treated with LD (p<0.001). Postoperative proptosis was 19.7±2.5 mm, and there was no statistically significant difference between the two groups (Table 4).

Table 4.

Surgical Outcomes According to the Technique Used

Surgical technique	Patients (orbits)	Preoperative proptosis [mean±SD (range), in mm] ^a	Postoperative proptosis [mean±SD (range), in mm] ^b	Proptosis reduction [mean±SD (range), in mm] ^a
MLD	197 (388)	25.0±2.4 (18–32)	19.7±2.6 (14–26)	5.7±2.2 (1–11)
LD	50 (97)	23.5±1.8 (19–26)	19.5±1.8 (15–23)	4.0±1.6 (1–8)
Total	247 (485)	25.0±2.4 (18–32)	19.7±2.5 (14–26)	5.4±2.2 (1–11)

A statistically significant difference between the two groups in terms of preoperative proptosis (^a p<0.001) can be observed. Postoperative proptosis was similar in both groups (^b p=0.4).

Patients' satisfaction

After a mean of 6 years following OD (range 2–9 years), we interviewed 192 out of 247 patients by telephone to ascertain their satisfaction. Both recovery of facial appearance and ocular function were evaluated using a semi-quantitative score as follows: 0=insufficient, 1=sufficient, 2=good, 3=excellent. More than 85% of patients reported good to excellent postoperative satisfaction for facial appearance and ocular function (see Table 5 for details). No significant difference was observed between the two surgical techniques in terms of patient satisfaction.

Table 5.

Patients' Satisfaction Regarding Appearance and Ocular Function

Level of satisfaction	n	%
Recovery of appearance
Insufficient	5	2.6
Sufficient	23	12.0
Good	81	42.2
Excellent	83	43.2
Recovery of ocular function
Insufficient	4	2.1
Sufficient	16	8.3
Good	77	40.1
Excellent	95	49.5

Patients were interviewed by telephone call after a mean of 6 years after orbital surgery (range 2–9 years).

Other side effects included hypoesthesia/anaesthesia of the infraorbital nerve (36 patients, 18.8%), persistent orbital pain (5 patients, 2.6%), and sinusitis related to the endonasal approach (3 patients on 161 operated on endonasally, 1.9%).

Discussion

GO is a relatively rare syndrome that, in patients with severe manifestations, can be disfiguring and sight threatening (3,4). The management of GO can be challenging and requires a multidisciplinary approach (3). Due to the anatomical structure of the orbit, expansion of the orbital cavity obtained by removing one or more orbital walls is a very effective manner of increasing the orbital space. After OD, prolapse of intraorbital tissues into the expanded volume results in reduction of proptosis, decreased optic nerve compression, and total or partial resolution of orbital congestion with the improvement of ocular symptoms (12).

OD is the mainstay of surgical treatment of GO. Several surgical approaches have been proposed during the last 20 years, and many studies have been published on the topic. For these reasons, these studies are not easily comparable, mainly because of the differences in preoperative patients' conditions and surgical approaches.

Although safe, OD is not free from postoperative complications. Postoperative diplopia is the main complication in terms of prevalence and worsening of the quality of life. Previous studies showed a wide range of the prevalence of postoperative diplopia depending on the surgeon's experience and the surgical technique (5,9 –12,15,16). Goldberg et al. compared their results after balanced medial plus lateral wall versus lateral wall only decompression (9). They observed that new-onset, persistent postoperative strabismus occurred in 33% of patients after balanced medial plus lateral wall decompression and in 7% of patients after lateral wall decompression. Graham et al. reported a rate of new-onset postoperative diplopia of 10% in a series of 40 patients operated by medial and lateral wall decompression (10). This series included patients with pre-existing diplopia in primary gaze. Ben Simon et al. concluded that deep lateral wall OD does not have significant effects on horizontal and vertical deviations (17). In a previous study, we observed an overall rate of new-onset or worsening of postoperative diplopia of 20% after MLD (11). In addition, we did not observe postoperative diplopia in primary gaze after LD in 39 patients, with only 8% of patients with new-onset diplopia in secondary gaze (12).

In our study we found that the most important feature affecting postoperative diplopia in primary gaze was the presence of preoperative diplopia in secondary gaze. This phenomenon can be mainly caused by a reduction of ocular version in the upgaze due to a fibrosis (with consequent restriction) of the inferior rectus muscle. We did not find correlations between postoperative diplopia in primary gaze and sex, age, year of surgery, severity of preoperative proptosis, extent of decompression (evaluated as proptosis reduction), and asymmetric presentation of the disease (evaluated as the difference in proptosis between the two eyes). Comparing the two surgical techniques (MLD vs. LD) in patients with preoperative diplopia, we did not observe a significant difference in the occurrence of diplopia in primary gaze. In contrast, in patients without preoperative diplopia, we observed that those treated with LD had a significantly lower prevalence of postoperative diplopia in primary gaze. This observation can be explained at least in part by the anatomical shape of the lateral wall of the orbit, which is almost posterior to the globe. Thus, displacement of the orbital contents toward the lateral wall after LD is limited. On the other hand, removal of the medial wall with manipulation of the orbital fat causes an infero-medial shifting of the orbital contents that can increase the risk of postoperative diplopia in primary gaze. Interestingly, a relevant finding is that in patients with preoperative diplopia in secondary gaze, the impairment of extraocular muscle motility did not depend on the surgical technique in terms of postoperative diplopia in primary gaze.

A recent review indicated that there is a wide range of exophthalmos reduction with any of the surgical techniques (18). This can be explained by the surgeon's skill and by the fact that surgical techniques are not standardized, and also because of anatomical differences. In our series, mean values for postoperative proptosis overlapped, regardless of the surgical techniques. These data are, of course, obtained starting from various values of preoperative proptosis in the two groups and, therefore, resulting in different reductions in proptosis using different techniques (LD resulted in a mean proptosis reduction of 4.1 mm, whereas MLD resulted in a reduction of 5.7 mm), thereby demonstrating that the choice between the two surgical techniques was adequate according to the preoperative values of proptosis. Looking at the patient's subjective evaluation after OD, it is important to emphasize that more than 85% of subjects in both groups experienced good to excellent satisfaction in terms of recovery of function and cosmetic outcomes. The finding that the rate of satisfaction overlapped in the two groups (with a different incidence of postoperative diplopia) may seem unexpected; but it should be emphasized that all of the patients were interviewed several years after their OD for functional and cosmetic outcomes (and after the surgical squint in those who had diplopia in the primary gaze). Therefore, at the moment of the interview, the two groups were relatively homogenous in terms of residual proptosis and motility of the extraocular muscles. This may explain the overlap in terms of the rate of satisfaction between the two groups.

The most frequent surgical complication observed after the diplopia in the primary gaze was a sensory defect of the infraorbital region. This was probably due to the overheating of the infraorbital nerve during the removal of the lateral wall in proximity of the inferior orbital fissure caused by drilling of the thick sphenoid and zygomatic bones.

In conclusion, patients with preoperative diplopia in the secondary gaze are more prone to develop diplopia in the primary gaze after OD, independently of the surgical technique. Among patients without preoperative diplopia, LD treatment seems to prevent postoperative diplopia in the primary gaze, compared with patients treated with MLD, although the use of the latter technique allows a higher reduction of proptosis. Therefore, to balance the risk of postoperative diplopia with the best rehabilitative goal, the planning of OD in patients with GO should take into account the presence of preoperative diplopia in the secondary gaze and the severity of proptosis. This aspect should be considered during the preoperative discussion with the patient about the risk/benefit ratio.

OD surgery is well tolerated; patients are usually satisfied with regard to the restoration of facial appearance, and postoperative ocular function and postoperative diplopia in the primary gaze can be corrected by squint surgery.

Although retrospective, this study has several points of strength. One is that it was performed in a very large number of patients treated homogeneously in a single center by a single surgeon, and the follow-up of the study cohort was performed by the same ophthalmologist. However, additional, prospective, randomized studies are needed to confirm our findings.

Footnotes

Disclosure Statement

The authors declare that no competing financial interests exist.

References

Bahn

. 2010. Graves' ophthalmopathy. N Engl J Med, 362:726–738.

Smith

. 2010. Pathogenesis of Graves' orbitopathy: a 2010 update. J Endocrinol Invest, 33:414–421.

Bartalena

, Baldeschi

, Dickinson

, Eckstein

, Kendall-Taylor

, Marcocci

, Mourits

, Perros

, Boboridis

, Boschi

, Currò

, Daumerie

, Kahaly

, Krassas

, Lane

, Lazarus

, Marinò

, Nardi

, Neoh

, Orgiazzi

, Pearce

, Pinchera

, Pitz

, Salvi

, Sivelli

, Stahl

, von Arx

, Wiersinga

. 2008. Consensus statement of the European group on Graves' orbitopathy (EUGOGO) on management of Graves' orbitopathy. Thyroid, 18:333–346.

Bartalena

, Tanda

. 2009. Clinical practice. Graves' ophthalmopathy. N Engl J Med, 360:994–1001.

Garrity

, Fatourechi

, Bergstralh

, Bartley

, Beatty

, DeSanto

, Gorman

. 1993. Results of transantral orbital decompression in 428 patients with severe Graves' ophthalmopathy. Am J Ophthalmol, 116:533–547.

Walsh

, Ogura

. 1957. Transantral orbital decompression for malignant exophthalmos. Laryngoscope, 67:544–549.

Kennedy

, Goodstein

, Miller

, Zinreich

. 1990. Endoscopic transnasal orbital decompression. Arch Otolaryngol Head Neck Surg, 116:275–282.

Mourits

PhM

, Koornneef

, Wiersinga

, Prummel

, Berghout

, van der Gaag

. 1990. Orbital decompression for Graves' ophthalmopathy by inferomedial, by inferomedial plus lateral, and by coronal approach. Ophthalmology, 97:637–641.

Goldberg

, Perry

, Hortaleza

, Tong

. 2000. Strabismus after balanced medial plus lateral wall versus lateral wall only orbital decompression for dysthyroid orbitopathy. Ophthal Plast Reconstr Surg, 16:271–277.

10.

Graham

, Brown

, Carter

, Song

, Nerad

. 2003. Medial and lateral orbital wall surgery for balanced decompression in thyroid eye disease. Laryngoscope, 113:1206–1209.

11.

Sellari-Franceschini

, Berrettini

, Santoro

, Nardi

, Mazzeo

, Bartalena

, Mazzi

, Tanda

, Marcocci

, Pinchera

. 2005. Orbital decompression in Graves' ophthalmopathy by medial and lateral wall removal. Otolaryngol Head Neck Surg, 133:185–189.

12.

Sellari-Franceschini

, Lenzi

, Santoro

, Muscatello

, Rocchi

, Altea

, Nardi

, Megna

, Marcocci

. 2010. Lateral wall orbital decompression in Graves' orbitopathy. Int J Oral Maxillofac Surg, 39:16–20.

13.

Olivari

. 1991. Transpalpebral decompression of endocrine ophthalmopathy (Graves' disease) by removal of intraorbital fat: experience with 147 operations over 5 years. Plast Reconstr Surg, 87:627–643.

14.

Sellari-Franceschini

, Muscatello

, Seccia

, Lenzi

, Santoro

, Nardi

, Mazzi

, Pinchera

, Marcocci

. 2008. Reasons for revision surgery after orbital decompression for Graves' orbitopathy. Clin Ophthalmol, 2:283–290.

15.

Sellari-Franceschini

, Berrettini

, Forli

, Bartalena

, Marcocci

, Tanda

, Nardi

, Lepri

, Pinchera

. 1999. Orbital decompression in Graves' disease: comparison of techniques. Acta Otolaryngol Ital, 19:307–314.

16.

Paridaens

, Lie

, Grootendorst

, van den Bosch

. 2006. Efficacy and side effects of “swinging eyelid” orbital decompression in Graves' orbitopathy: a proposal for standardized evaluation of diplopia. Eye (Lond), 20:154–162.

17.

Ben Simon

, Syed

, Lee

, Wang

, Schwarcz

, McCann

, Goldberg

. 2006. Strabismus after deep lateral wall orbital decompression in thyroid-related orbitopathy patients using automated Hess screen. Ophthalmology, 113:1050–1055.

18.

Borumandi

, Hammer

, Kamer

, von Arx

. 2011. How predictable is exophthalmos reduction in Graves' orbitopathy? A review of the literature. Br J Ophthalmol, 95:1625–1630.